Date Posted: 10/21/2016

A recent paper on fairness and bias in algorithmic predictions by first-year CS PhD student Manish Raghavan, CIS faculty member Jon Kleinberg, and Harvard economist (and Cornell alum) Sendhil Mullainathan, was discussed by the Washington Post.

Their work sheds light on recent controversies about the potential for bias in algorithmic risk tools used in the criminal justice system. The paper establishes inherent trade-offs between competing definitions of algorithmic fairness: except in highly constrained cases, these definitions cannot all be simultaneously satisfied.